AI Textile Mill Onboarding: Stanford Capstone with SAP
Initial conversations with SAP, project mission:
How can enterprise supply chain systems like SAP Business Network better support the relationship-driven, fast-moving realities of fashion sourcing while designing for transparency and resilience.
Research
My team of four and I interviewed 14+ supply chain
managers and materials specialists at fashion brands like Nike and Kith, salespeople at textile mills across the world, and supply chain and sourcing professionals in related industries like chemicals
Process photo of my team mate Amelie and I working on organizing and synthesizing our interviews.
The problem we uncovered
Fashion sourcing teams are under pressure to vet new textile suppliers, driven by tariff shifts and tightening sustainability requirements. Yet the primary tool for doing so, the fabric detail sheet (FDS), consistently fails on both sides: buyers rarely receive complete information back, and suppliers find the process time-intensive and error-prone.
Materials specialists managing textile mill discovery need to quickly vet new suppliers without the late-stage surprises (missed collection windows, wholesale penalties) that stem from incomplete information.
Suppliers already maintain internal FDS documentation with most of the information buyers need, but right now they are forced to manually re-enter it into a different format for each buyer, leading to them often submitting incomplete or inaccurate information.
Solution: Supply Connect
We designed an AI-powered onboarding layer for an enterprise software platform. Suppliers upload their internal fabric documentation once. AI maps supplier information directly into buyer FDS formats
After presenting and validating our idea during our capstone presentation, I personally prototyped a functional first version of the FDS re-mapping engine for a supplier, you can access it here (Intended for computer, the site is password protected to prevent spam: Username: Leo, Password: Supplyconnect).
